Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Clean way to run 3 phase induction motor

  1. Jan 22, 2016 #1
    Say you wanted to run an induction motor from a battery without putting dirty harmonics into it.
    Would a really clean way to do it, to be by using a higher phase number inverter, then transformer, then into the motor?
    For instance, a 6 phase inverter, into a 6 phase to 3 phase transformer, then into the induction motor?
  2. jcsd
  3. Jan 22, 2016 #2


    Staff: Mentor

    Yes, those things you said will get you a cleaner AC voltage.

    The absolute cleanest would be to use a DC motor to drive an AC generator to drive the AC induction motor. But that is probably too much trouble.

    It is pretty unusual to run an induction motor from a battery. Have you compared the battery energy and power delivery specs compared to the motor's needs?
  4. Jan 22, 2016 #3
    Nice reply.
    No, purely theoretical, I am aware of the difficulty storing electricity. I was just thinking about Tesla cars, hence my unusual question.
  5. Jan 22, 2016 #4
    Is it a requirement the motor be a 3ph AC motor? How about a brushless DC (BLDC) motor? I believe they have better torque than 3ph AC at low speeds, and can provide position feedback even.
  6. Jan 23, 2016 #5
    As I said, purely theoretical, I just know that Tesla uses induction motors, I assume 3 phase.
  7. Jan 25, 2016 #6
    You can build an analog inverter with VERY clean waveform - but they are inefficient. When Using VFD the trade off is high switching frequency = better waveform = higher losses. So a 6 phase inverter + transformer at 4Khz Switching would look the same as a three ph inverter at 8khz. The additional looses due to 2 x the number of devices and the transformer - make the basic 3ph inverter the better choice.

    Some standard industrial VFDs allow DC direct input and adjustment of the switching frequency - but the DC input would need to be higher than the motor voltage by sqrt(2).

    A good example is large UPS system where power quality is more important than typical motor drive - here they are often using mulilevel inverter topology. These are actually more efficient at the higher frequencies, but the downside is the system cost and complexity. I will not bore you wiht the details but this Ap Note may,,,, haha
  8. Jan 25, 2016 #7


    User Avatar
    Science Advisor

  9. Jan 25, 2016 #8


    User Avatar
    Science Advisor

    Many EV's (electric vehicles) use AC motors so there's a wide range of commercial inverters available. They are often water cooled and have good efficiency (~98%). Tesla may have a proprietary design but it's probably not wildly different to the rest.
    DIY EV enthusiasts have also developed their own open source inverters so you can find full schematics & parts list for free online.
  10. Jan 26, 2016 #9
    You may think I have a hide for asking, but what is the go with multilevel and cascaded inverters? I only learnt about 'inverters' when I took a power electronics course, I'm happy to read through the application note and topology when I get a chance, but a quick explanation of what they are (maybe their advantages) would make it seem less daunting. Thanks
    Yeah I would be after something with decent efficiency, as you said isn't it that switching frequency key to having a cleaner supply, push the harmonic up high with high switching frequency and then filter it or something?
    Those numbers you threw up, was that a rough guess from your experience, or were you just making a point?
    I see that you could lose a few percent efficiency with the transformer, but say you were just using a 6 phase vs a 3 phase inverter, how would they be so similar in loss for comparable quality of output (regarding the necessary switching frequencies)?

    Thanks for the tip, but do you have any favourites in particular of your own you'd suggest?

    Thanks for that, I had a very quick look at it, was there anything specifically you were drawing my attention to?
  11. Jan 26, 2016 #10
    My 2kw VFD I use for my lathe (Allen Bradly Powerflex 50) has an adjustable switching frequency from 2khz to 20 khz... You lose some power output the higher you go (derating) but get cleaner power with no audible noise.
  12. Jan 26, 2016 #11
    You mean the output power has to be less than 2kW if you go up to 20kHz switching frequency?
    What do you mean by 'derating'?
    I used the powerflex 525 before, just V/f control, if I was to have done space vector control, or some other method, how would that have worked? I mean I learnt park and clarke transforms once before, but I don't see how I'd implement that into using/controlling the actual drive?
  13. Jan 26, 2016 #12


    User Avatar
    Science Advisor

    No, I linked to it, since it had explanations of what a variable speed motor controller is and what is commercially available.
  14. Jan 26, 2016 #13
    Yeah ok, I just thought you were getting at something particularly relevant to this thread.
  15. Jan 28, 2016 #14
    As for my inverter numbers --

    In the inverter we have 2 kinds of losses, conduction and switching - some conduction losses are resistive ( so paralleling helps) some are due to Vf - so paralleling makes no difference, and switching is generally a fixed amount for each pulse ( under fixed conditions, and varies over a waveform cycle).

    In a 6 pulse - lets say you use Phase 1 and 2 to create the Phase A in the 3 phase output ( after the transformer) - the 6 pulse PWM firing would be shifted Phase 1 from phase 2 so the output waveform will have effectively 2 x the switching frequency as these two phases are ADDed in the transformer. ( really there are many ways to do this) - so to make the same quality waveform in 3 phase inverter you need to switch at the 2 X speed. ( Note - since MOSFETs have much lower switching losses than IGBTs - the impact is less, but getting large power / voltage in MOSFETs is difficult, so large power ( approx 1 KW or > 240VAC) uses IGBTs and pays more of a price in switching.

    So you want to have the same quality signal using a 6 phase as a 3 phase - the 6 phase will use 6 half bridges (12 switching devices) and the 3 phase 3 half/6 devices ( star connected). The 3 phase will have to switch 2 as fast to meet the same waveform as the 6 phase, and the devices will see more current, they will have higher losses for the inverter section. The 6 phase will have a little extra filtering due to a transformer, but the filtering (higher frequencies) increase the losses in the transformer, vs a pure 60Hz sine. So generally the cost and complexity of the 6 phase is not worth the effort, except is some unusual situations ( like using a transformer to step up and make a Medium Voltage drive) -

    The miltilevel reduces the switching losses by witching more "levels" of DC to make the ac waveform - in low voltage the majority is referred to a three level - the DC link is set up with 2 caps in series and you have 3 DC levels to switch to the output, think "+ to N to - " vs 2 levels for a standard inverter, just + and -.

    So as the AC output is created - part of the waveform is switched using 1/2 of the total DC voltage, then for the middle of the AC waveform the full DC link voltage is used - but during this part of the cycle the switches controlling the middle half of the DC V( I really have no other way to describe it) are left ON, not pulsed, and only the switches controlling the full DC V are pulsed.

    It may take a while to see the benefit here. - But each device is switching at a lower voltage, better duty cycle, and for less time over the cycle. Essentially it cuts the switching losses in half.

    The Cascade types- another versions of multilevel are of course more complex, and typically used in Medium Voltage ( MW + ) - and you can get many levels. Many need front end transformers to generate independent DC voltages - they are not cheap because of this. So while they do work, many cases it is cheaper to use Low Voltage and parallel a lot of low voltage and use a single ( almost standard type) transformer to step up the V.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook